Minutes of the Geant4 Technical Steering Board Meeting no 14 27 September 2001

Submitted by Anonymous (not verified) on 
(3rd Draft)

Attendance:

CERN: Katsuya Amako, John Apostolakis (chairman), Pedro Arce, Gabriele Cosmo, Gunter Folger, Vladimir Ivantchenko, Jurgen Knobloch, Michel Maire, Maria Grazia Pia, Florence Ranjard, Sergei Sadilov, Matthias Schroeder, Peter Truscott

SLAC: Makoto Asai, Dennis Wright

Ecole Politechnique : Marc Verderi

KEK: Youhei Morita, Takashi Sasaki

TRIUMF: Peter Gumplinger

Minutes: F.Ranjard - 18 October 01

Regular Items

  • New Geant4 collaborators:
    • SLAC
      • Tony Johnson - Geant4 interfaces
      • Joseph Perl - WIRED event display for Geant4
      • Doug Smith - HyperNews work
    • PPARC
      • Colin Ferguson / Department of Physics & Astronomy, Southampton University

        • possible working groups (geometry, low-E e.m) to be discussed.
  • Minutes of last meeting: approved
  • Review of pending actions
    • 11.9 closed : support for Sun CC 4.2 will be stopped.
    • 11.10 closed
    • 11.13 closed
    • 11.14 see 10.4
    • status of remaining actions can be consulted in the appendix.

Review Report (J.Apostolakis)

nothing new.

Documentation Review (D.Wright)

  • the global review is being done: comments have been sent, or will be sent in the next days, to authors.
  • Authors are invited to implement the suggested changes by 14 November.
  • In case authors find difficult to implement the changes, that should be reported to the TSB.

Geant4 Paper (J.Allison)

  • the new schedule proposed by J.Allison is agreed on .

    • action 14.1 J.Allison to give a more precise schedule to report comments on the paper.

Preparation for next release (G.Cosmo)

  • ending of STL migration: hadronic physics is a problem
    • completion of the migration is strongly recommended but if it should delayed the release, it will be a concern.

      • action 14.2 J.Apostolakis will contact category coordinators to understand any reason of the delay.
  • cleanup of all warnings due to new gcc compiler
    • fix problems in hadronic category which exist with both compilers.
  • category coordinators should not wait till the last moment to commit their changes.
  • documentation should be there by 1 December.
  • send documentation of new features to the Documentation coordinator.

    • action 14.3 G.Cosmo to send a reminder with schedule to Category coordinators.

New features to be included into geant4.4.0

  • Run & Event category (M.Asai)
    • HepMC interface
    • Stacking manager will allow the user to increase thenumber of stacks without modifying G4 source code. This feature will help the user to control the priorities of processing tracks in much more flexible manner. This feature has already been implemented and passed through the STT.
  • Geometry (G.Cosmo)
    • Magnetic field per volume
    • 3 extended examples
    • some commands to ease change of geometry.
    • reflection of solids
    • integration of BaBar tool to check geometry
    • migration to STL

      • action 14.4 J.Apostolakis to look at the magnetic field problem in vacuum and in dense material ( a G4Exception instead of a core dump does not solve the problem)
  • hadronic physics
    • 3 new features (no detail in the absence of J.P.Wellish)
  • electromagnetic physics (M.Maire)
    • new X-scattering model presented in Genova.
    • generation of sub-cutoff delta-ray in G4MuIonisation.
    • remove production cuts in photoelectric, Compton, gamma conversion, positron annihilation effects.
    • first implementation of Store/Retrieve physics tables in all processes
    • The em processes using the algorithm called 'integral approach' to compute the step length are declared obsolete and will be no longer supported (G4IeIonisation, ..etc..).
  • low-energy electromagnetic physics (M.G.Pia)
  • particles & track (H.Kurasige)
    • reduction of initialization time: finish implementation of store/retreive physics table for eack w.m.precess.
    • cut per material/volume/region: make the necessary changes in the code to allow cut per material. Cut per region could achieved by setting different materials for different regions.

      • these changes imply changes in several categories (e.m and low-E e.m.)
  • Visualization (J.Allison)
    • trajectory in magnetic field are badly visualized. Visualization needs more information along the step. No improvement yet in next release.
  • Analysis
    • being redesign

      • action 14.5 J.Apostolakis to nominate category coordinator to make plans in view of AIDA evolution.
  • Processes & Materials (M.Verderi)
    • parametrisation models are not part of the testing procedure.

      • action 14.6 S.Sadilov to obtain tests to be included in the testing procedure.

Technical topics

  • cuts per material/volume/region (H.Kurasige)
    • P.Arce has a draft implementation of cuts per material, however there were concerns about whether this should be added in next release only to be removed in a later version early next year when cuts per region becomes available; proper design methodology and whether the impact of such changes on the processes has been considered carefully.

      • action 14.11 J.Apostolakis to consider the issues and propose a timescale for implementation by mid October following the consultation with people like Pedro and Hisaya.
  • STL migration of Materials category code (M.Verderi)
    • requires the migration of Hadronic processes code.
  • W2000 sharable libraries (G.Cosmo)
      • action 14.7 code writers are asked to move the body of static accessors to static storages from header files to implementation files.

Software Process

  • update on global context activities (G.Cosmo)
    • many diagrams are missing. CC are remembered that diagrams are part of the documentation which should be committed by 1 December 01.
  • Ignominy
    • the tool has been exercised on geant4. It shows that some categories have many dependencies on other categories.
    • the tool is too "young" to be used by default in the testing procedure.

      • action 14.8 G.Cosmo to organize a presentation of the tool to the next CC meeting.

Report from Experiments and Institutes

  • BaBar
    • about 10M events have been produced with very few problems reported to Geant4 team.
    • withdraw the Geant3 based program in 6 months
  • PPARC
    • (should try to make a link to J.Allison mail - Flr)
  • LHCb
    • W2000 server is ready with possibility of remote login but without afs/
    • Try now to build geant4 on \cern.ch\dfs.
  • ESA
  • HARP
    • simulation framework is ready but not yet in production
  • CMS
    • OSCAR is ready and complete.
    • Test OSCAR against Geant3 based program and test beam data
    • expect OSCAR to replace Geant3 based program in about a year from now.
  • ATLAS
    • e.m physics should be tested by end '01
    • hadronic physics should be tested by Summer '02
    • use the GOOFY framework by the end '01
  • SLAC
    • will host a Geant4 user's workshop from February 18th, 2002 to February 22nd, 2002. Though this will be a domestic workshop, participants from the world are welcome.

A.O.B

  • action 14.9 J.Apostolakis to get a draft of 2002 milestones for the next TSB meeting
  • action 14.10 S.Sadilov will provide a prototype based on Bugzilla implementing the template for user requirements.